This invention relates to apparatus for measuring length or angle, which include a data reading apparatus using a storage effect type sensor such as a charge coupled device (abbreviated CCD).
The measurement of the translational displacement or angle of rotation of an object from a reference point under remote control may be generally made by providing a scale on the object, effecting zero resetting at a position constituting the reference point of the scale, optically or electromagnetically counting the translational displacement or rotational angle and displaying the result, for instance on a display unit.
However, in this method, which is based upon a commonly termed incremental system for measuring the relative movement of the scale and sensor from an electric signal generated at the time of the movement, it is necessary to detect the reference point.
To overcome this inconvenience, an absolute system where the scale is encoded has been proposed. This system uses a scale, which is graduated with codes of lenghts from a reference point, so that a given length from the reference point can be obtained by reading the scale without need of detecting the reference point.
The feature of this absolute system that it is not necessary to detect the reference point leads to a feature that the system permits intermittent or sampling measurements. With the incremental system, which depends for measurement upon the count after the zero resetting, the sampling measurements are difficult to make because of the fact that the reading system and processing system that constitute the incremental system has to be always operating.
With regard to the construction of the absolute system, however, compared to the incremental system, which comprises a reading system dealing with a signal from a few number of sensors (a single sensor in the extreme case) and a processing system mainly constituted by a counter section, the absolute system has to use a great number of sensors for the reading of the scale, and also the shape of the sensor is restricted since the size of the sensor corresponds to the quantity of signal. Therefore, in order to be able to obtain a sufficient resolution, it is necessary to permit fine measurement with some or other means, and this dictates many disadvantageous requirements to the reading system and processing system of the absolute system.
As a solution to these problems in the absolute system, it is thought to use a storage effect type sensor. The storage effect type sensor has a number of sensor elements and outputs the stored signal as a serial signal, so that the reading system and processing system can be simplified.
Now, the absolute system using a storage effect type sensor will be described.
The storage effect type sensor has a construction of, for example, a charge coupled device (CCD) having a photoelement section, a gate section and a shift register section. The photoelement section has a number of light receiving elements arranged in a row and each having a function of storing charge introduced in correspondence to the intensity of the incident light. The shift register section has bits corresponding in number to the number of light receiving elements in the photoelement section. The charges stored in the individual light receiving elements are simultaneously transferred through the gate section to the respective bits, and when a scanning pulse signal is given the charges transferred to the individual bits are sequentially provided as a time series output signal. In the storage effect type sensor having the above construction, the width of each light receiving element constituting the photoelement section may usually be set to the order of ten and several microns, so that it is possible to obtain the reading of the scale and also subdivide or interpolate each graduation.
As shown in FIG. 1, a storage effect type sensor, for instance a CCD Y, having a photoelement section F, gate section G and a shift register section R, is provided to face one side of a scale C, in which bright (or transparent) sections A and dark (or non-transparent) sections B are alternately arranged along a straight line. When a parallel light beam P having a width X is projected onto the other side of the scale C, charges are stored in the light receiving elements in the photoelement section F corresponding to bright sections A of the scale in accordance with the incident light intensity. At this time, by supplying a gate pulse s and also a reset pulse signal Z and scanning pulses .phi..sub.1 and .phi..sub.2 having one-half the frequency of the pulse signal Z and 180.degree. out of phase from each other, as shown in FIG. 2, to the CCD Y, an output CO as shown in (d) in FIG. 2 is obtained from a differential amplifier H. In this example, the width of the bright sections a is set to correspond to three light receiving elements in the photoelement section F and three light receiving sections accurately face to each bright section A, so that the differential amplifier H provides a time series pulse signal including the corresponding three consecutive pulses at a time in accordance with the aforementioned correspondence relation. Thus, by appropriately combining a scale in which the length from a reference point is encoded in terms of a bright and dark bit pattern and a storage effect type sensor, it is possible to obtain measurement of the distance, i.e., length, from the reference point of the scale with a precision factor of ten and several microns or less. FIG. 3 shows a specific example of the combination of the two. Here, the length of the scale C is divided into a plurality of blocks of the same length. In the Figure, block N.sub.p and N.sub.p-1 are shown. Taking the block N.sub.p, for instance, a marker M is formed at the left end of this block in the form of a bright section having a width corresponding to four light receiving elements in the photoelement section F of the CCD Y, each light receiving element being given a unit width. The block N.sub.p-1 adjacent to the block N.sub.p has a block stop marker D formed at its right end in the form of a dark section having a width corresponding to four light receiving elements of the CCD Y. Of the marker M, which corresponds in width to four elements, a portion adjacent to the block stop marker D and having a width corresponding to one element is made as a market bit I.sub.m. On the right side of the marker M, dark sections K.sub.1, K.sub.2, K.sub.3, . . . having a width corresponding to two elements and spaced apart from one another by an interval corresponding to two elements are formed. Further, address sections P.sub.1, P.sub.2, P.sub.3, . . . each corresponding in width to two elements are provided between adjacent dark sections, and binary codes 2.sup.0, 2.sup.1, 2.sup.2, . . . are allotted to the respective address sections P.sub.1, P.sub.2, P.sub.3, . . . In the illustrated example, the address sections P.sub.1 and P.sub.3 are formed as bright sections, while the other address sections are formed as dark sections. In the block N.sub.p-1, similar address sections are also formed, but in this block the values of the binary coded address sections allotted are those in the block N.sub.p minus 1.
Using the scale C and CCD Y having the arrangement as described above, the distance from the reference point of the scale C can be measured in the following way.
The individual light receiving elements constituting the photoelement section F of the CCD Y are given respective bit numbers as shown, and a particular bit is made as an index bit I.sub.n. The scale C is then irradiated with a parallel light beam with a width X greater than that of two blocks, and the state of the CCD Y is examined. In the first place, the position of the photoelement section corresponding to the marker M and block stop marker D, individually corresponding in width to four elements, is determined, and then the position corresponding to the marker bit I.sub.m is determined. In the illustrated case, the bit labeled 0 in the photoelement section of the CCD Y corresponds to the marker bit I.sub.m. Since the positional relation of the marker bit I.sub.m to the address sections P.sub.1, P.sub.2, P.sub.3, . . . is clear, the position of a given address section of the photoelement section of the CCD Y can be known from this positional relation and the bit 0 corresponding to the marker bit I.sub.m. At this time, whether the bits labeled 6 and 7 , 10 and 11 , 14 and 15 , etc. are "on", i.e., receiving light, is detected. In the instant case, the bits 6 and 7 are "on", that is, the 2.sup.0 place order bit is "on". Also, the bits 10 and 11 , i.e., the 2.sup.1 place order bit, is "off". Likewise, the bits 14 and 15 , i.e., the 2.sup.2 place order bit, is "on", and the bits 18 and 19 , i.e., the 2.sup.3 place order bit, is "off". Consequently, the absolute address of the block N.sub.p is 2.sup.2 +2.sup.0 =5, indicating that the block N.sub.p is a fifth one from the reference point of the scale. If one block consists of 26 bits as shown in the Figure and the width of one bit is 15 .mu.m, the width of one block is 15 .mu.m.times.26 =0.39 mm. Accordingly, it can be known that the block N.sub. p is moved by 0.39 mm.times.5=1.95 mm from the reference point. Subsequently, for obtaining the accurate distance, the bits from the index bit I.sub.n to the bit 0 corresponding to the marker bit I.sub.m are counted. In the instant example there are 5 bits, and hence the accurate distance is 1.95 mm-15 .mu.m.times.5=1.875 mm.
In general, denoting the number of bits from the index bit I.sub.n to the marker bit I.sub.m by N.sub.c, the address of the block by ACD, the number of elements in the photoelement section F corresponding to one block length by N.sub.s and the width of one element by m .mu.m, the measurement value E is EQU E=m(N.sub.s .multidot.ACD-N.sub.c) (.mu.m).
While this equation applies in case of the measurement of the address of the block, whose marker is on the right side of the index bit I.sub.n in FIG. 3, in the case of the measurement of the address of a block whose marker is on the left side of the index bit I.sub.n, the measurement is given as EQU E=m(N.sub.s .multidot.ACD+N.sub.c).
Further, when increasing the precision of measurement within the width of one bit of the element of the photoelement section, a vernier scale section may be provided between the marker section and address sections of each block such that the number of photoelement section elements in the same length as the vernier scale section is N.sub.B +1 where N.sub.B is the number of bits of the vernier section, and the state of overlap between the scale side vernier scale section bits and the photoelement section bits may be examined (to determine the position of the bits in right overlap, for instance).
By combining the scale which has an absolute address based upon a bright and dark bit pattern and a storage effect type sensor, it is possible to measure the distance, i.e., length, with high precision without need of detecting the reference point, and this method can also be directly applied to the measurement of angle.
However, if it is intended to make actual measurement of the length or angle by using the storage effect type sensor in the manner as described, the measurement contains a large error due to the following problems.
Taking now the case of measurement by storing charge in the photoelement section F in the CCD Y shown in FIG. 1 in the period T.sub.1 shown in FIG. 4, transferring the charge at the instant of end of the period T.sub.1 to the shift register R by the bit parallel method with the supply of the gate pulse S as shown in (a) in FIG. 4 and reading out the content of the shift register R in a period T.sub.2, the output CO of the operational amplifier H is usually as shown in (e) in FIG. 4, having a certain slope with respect to a reference level. As is shown, at the start of the reading from the shift register R the reference level is already drifted by V.sub.a from an ideal level, for instance 0 level, and at the end of the reading it is further drifted by V.sub.b from V.sub.a. Such variations of the reference level result from the dark current in the photoelement section F and shift register section R. V.sub.a is a voltage corresponding to the charge stored by the dark current in the photoelement section F, while V.sub.b is a voltage corresponding to the charge stored with the dark current in the shift register section R. The value of V.sub.a is proportional to the period T.sub.1, while the value of V.sub.b is proportional to the period T.sub.2. With such variations of the reference level, the signal V.sub.c read out also varies with respect to the aforementioned level, so that an erroneous operation of the system for processing the read-out signal is prone. Also, the values of V.sub.a and V.sub.b vary with temperature, and in an extreme case the reference level reaches the saturation level V.sub.D due to the dark current component, and in such a case the read-out of signal can no longer be obtained. Particularly, when it is intended to make signal processing as mentioned above with the combination of the scale C having bright and dark portions and CCD Y as shown in FIG. 3, with the aforementioned variation of the output signal level accurate measurement cannot be expected because the reading of the address and detection of the number of bits from the index bit to the marker bit have to be effected on the basis of the level of the output signal for each bit.
From the above ground, in the usual method using CCD, the frequency of the scanning pulse signals .phi..sub.1 and .phi..sub.2 supplied to the CCD Y is set to a sufficiently high value, for instance above 100 kHz, to reduce the values of V.sub.a and V.sub.b. However, where the CCD is used in a system processing output of the CCD, the output CO is required to be processed at least within 5 .mu.sec. To this end, an expensive high-speed A/D converter is required, and also it is necessary to use an expensive high-speed microcomputer for the processing of the output of the aforementioned high-speed A/D converter and control of the CCD. Further, since V.sub.a and V.sub.b are influenced by temperature even if the frequency of the scanning pulse signals is increased by adopting the aforementioned method, means for correcting V.sub.a and V.sub.b are required in order to ensure reliable signal processing operation and improve the measurement precision.
In a further aspect, where the storage effect type sensor, typically CCD, is used, although the measurement can be conveniently obtained when the scale C is stationary, when the scale C is moving, the position of the photoelement section illuminated by light through the bright portions of the scale C is instantaneously changed, that is, information of the state of light reception is recorded for each instant. Therefore, the maximum speed of the scale C at which the measurement can be made is limited to a comparatively low speed. If the maximum allowable speed is increased by some or other means, the condition for using is greatly restricted.
Insofar as the aforementioned various means are necessary for using the storage effect type sensor, the simplification of the data reading and processing systems sought as a preamble of the invention in using a storage effect type sensor, can not be achieved.